Allow removal to non edge bins with RemoveTimeBins (also rename to RemoveBins)

[Nick Draper]

Allow the removal of bins from the middle of the range. Add a parameter for interpolation (initially allow None and linear). If linear interpolation interpolate the value across the remaining bins.

If None then just remove it.

[Matt Clarke]

(In [1872]) First attempt at the new RemoveBins algorithm. Refs #373

[Matt Clarke]

(In [1874]) Refs #373. Some changes.

[Matt Clarke]

(In [1877]) Refs #373. Forgot to save it before commiting!

[Matt Clarke]

(In [1885]) Refs #373. Should hopefully not set all other values to zero!

[Matt Clarke]

(In [1895]) Refs #373. Testing buildserver.

[Nick Draper]

We need to alter the StartX and EndX to take the x bin values, not the indexes.

Just iterate over the X array until you find the first bin that would contain the value, and then crop until the bin that contains the EndX value.

[Matt Clarke]

(In [1913]) Refs #373. RemoveTimeBins now takes xvalues.

[Matt Clarke]

(In [1915]) Refs #373. Adjusted SConstruct

[Matt Clarke]

(In [1919]) Refs #373. Updated sendStatusEmail.py so it recognises "Failed 1 of 1 test"

[Russell Taylor]

(In [1924]) RemoveBins wasn't copying the spectrum numbers across to the output workspace. Re #373.

[Russell Taylor]

(In [1947]) RemoveBins can now take in an X range which is in different units to that of the input workspace. This is required for the SANS instruments. Also, when not interpolating, the counts in a bin that straddles the edge of the range is scaled proportionately to the portion of the bin that is within the range. Indices used for interpolation end points should be looked at at some point, and the tests should be expanded... Re #373.

[Russell Taylor]

(In [1948]) Mistaken check-in. Re #373.

[Russell Taylor]

(In [1950]) Fix RemoveBinsTest. Re #373.

[Russell Taylor]

At some point, the interpolation will need upgrading to, e.g., cubic spline.

[Stuart Campbell]

This ticket has been transferred to github issue ​1221